@oscharko-dev/keiko 0.2.0-beta.7 → 0.2.0-beta.9

package/node_modules/@oscharko-dev/keiko-workspace/dist/repoSearchPolicy.js

@@ -0,0 +1,342 @@
+const METADATA_FILENAMES = new Set([
+    "package.json",
+    "tsconfig.json",
+    "tsconfig.base.json",
+    "tsconfig.build.json",
+    "vite.config.ts",
+    "vite.config.js",
+    "vitest.config.ts",
+    "vitest.config.js",
+    "jest.config.ts",
+    "jest.config.js",
+    "playwright.config.ts",
+    "playwright.config.js",
+    "next.config.ts",
+    "next.config.js",
+    "eslint.config.ts",
+    "eslint.config.js",
+]);
+const OVERVIEW_FILENAMES = new Set([
+    "readme.md",
+    "agents.md",
+    "contributing.md",
+    "architecture.md",
+    "docs/architecture.md",
+    "docs/overview.md",
+    "docs/readme.md",
+]);
+const LOCKFILE_FILENAMES = new Set([
+    "bun.lock",
+    "bun.lockb",
+    "cargo.lock",
+    "composer.lock",
+    "gemfile.lock",
+    "package-lock.json",
+    "pnpm-lock.yaml",
+    "yarn.lock",
+]);
+const LOW_VALUE_SEGMENTS = new Set([
+    ".parcel-cache",
+    ".svelte-kit",
+    ".vercel",
+    "coverage",
+    "dist",
+    "generated",
+    "storybook-static",
+    "tmp",
+]);
+const LOW_VALUE_IGNORE_LINES = Object.freeze([
+    ".parcel-cache/",
+    ".svelte-kit/",
+    ".vercel/",
+    "coverage/",
+    "dist/",
+    "generated/",
+    "storybook-static/",
+    "tmp/",
+    "bun.lock",
+    "bun.lockb",
+    "cargo.lock",
+    "composer.lock",
+    "Gemfile.lock",
+    "package-lock.json",
+    "pnpm-lock.yaml",
+    "yarn.lock",
+]);
+const SOURCE_EXTENSIONS = new Set([
+    "cjs",
+    "cts",
+    "go",
+    "java",
+    "js",
+    "jsx",
+    "mjs",
+    "mts",
+    "py",
+    "rs",
+    "ts",
+    "tsx",
+    "vue",
+]);
+const DOC_EXTENSIONS = new Set(["adoc", "md", "mdx", "rst", "txt"]);
+const TEST_FILE_RE = /(?:^|[./_-])(?:test|spec|fixture|mock)s?(?:[./_-]|$)/iu;
+// Identifier tokens with a SINGLE greedy quantifier — no two overlapping `[A-Za-z0-9_$]*` runs
+// around `[A-Z]`, which is a polynomial-ReDoS shape on uncontrolled query text (CodeQL). The
+// camelCase test (an uppercase at position >= 1) is applied separately as a constant-width probe.
+const IDENTIFIER_TOKEN_RE = /\b[A-Za-z_$][A-Za-z0-9_$]*\b/gu;
+const HAS_INNER_UPPERCASE_RE = /[A-Za-z0-9_$][A-Z]/u;
+const TOKEN_RE = /[A-Za-z0-9_.-]{3,}/gu;
+function emptyBucketCounts() {
+    return {
+        "canonical-metadata": 0,
+        "overview-doc": 0,
+        "exact-path": 0,
+        "symbol-source": 0,
+        source: 0,
+        test: 0,
+        docs: 0,
+        lockfile: 0,
+        "low-value": 0,
+        other: 0,
+    };
+}
+function basename(scopePath) {
+    const index = scopePath.lastIndexOf("/");
+    return index >= 0 ? scopePath.slice(index + 1) : scopePath;
+}
+function extension(scopePath) {
+    const name = basename(scopePath).toLowerCase();
+    const dot = name.lastIndexOf(".");
+    return dot < 0 ? "" : name.slice(dot + 1);
+}
+function normalizedPath(scopePath) {
+    return scopePath.split("\\").join("/").toLowerCase();
+}
+function pathSegments(scopePath) {
+    return normalizedPath(scopePath)
+        .split("/")
+        .filter((segment) => segment.length > 0);
+}
+function isLockfile(scopePath) {
+    return LOCKFILE_FILENAMES.has(basename(scopePath).toLowerCase());
+}
+function hasLowValueSegment(scopePath) {
+    return pathSegments(scopePath).some((segment) => LOW_VALUE_SEGMENTS.has(segment));
+}
+function normalizedQueryTerms(query) {
+    const terms = new Set();
+    for (const match of query.text.matchAll(TOKEN_RE)) {
+        terms.add(match[0].toLowerCase());
+    }
+    for (const match of query.text.matchAll(IDENTIFIER_TOKEN_RE)) {
+        if (HAS_INNER_UPPERCASE_RE.test(match[0])) {
+            terms.add(match[0].toLowerCase());
+        }
+    }
+    return [...terms];
+}
+function bucketByPath(scopePath) {
+    const path = normalizedPath(scopePath);
+    const name = basename(path);
+    const ext = extension(path);
+    if (METADATA_FILENAMES.has(name)) {
+        return "canonical-metadata";
+    }
+    if (OVERVIEW_FILENAMES.has(path) || OVERVIEW_FILENAMES.has(name)) {
+        return "overview-doc";
+    }
+    if (isLockfile(path)) {
+        return "lockfile";
+    }
+    if (hasLowValueSegment(path)) {
+        return "low-value";
+    }
+    if (TEST_FILE_RE.test(path)) {
+        return "test";
+    }
+    if (SOURCE_EXTENSIONS.has(ext)) {
+        return "source";
+    }
+    if (DOC_EXTENSIONS.has(ext)) {
+        return "docs";
+    }
+    return "other";
+}
+function bucketScore(bucket, intent) {
+    if (intent === "project-metadata") {
+        return metadataBucketScore(bucket);
+    }
+    if (intent === "repository-overview") {
+        return overviewBucketScore(bucket);
+    }
+    if (intent === "targeted-code-search" || intent === "diagnostic-search") {
+        return targetedBucketScore(bucket);
+    }
+    return genericBucketScore(bucket);
+}
+function metadataBucketScore(bucket) {
+    const scores = {
+        "canonical-metadata": 100,
+        "overview-doc": 70,
+        "exact-path": 95,
+        "symbol-source": 80,
+        source: 45,
+        test: 40,
+        docs: 55,
+        lockfile: 15,
+        "low-value": 5,
+        other: 20,
+    };
+    return scores[bucket];
+}
+function overviewBucketScore(bucket) {
+    const scores = {
+        "canonical-metadata": 75,
+        "overview-doc": 100,
+        "exact-path": 95,
+        "symbol-source": 80,
+        source: 50,
+        test: 35,
+        docs: 70,
+        lockfile: 5,
+        "low-value": 5,
+        other: 20,
+    };
+    return scores[bucket];
+}
+function targetedBucketScore(bucket) {
+    const scores = {
+        "canonical-metadata": 65,
+        "overview-doc": 35,
+        "exact-path": 100,
+        "symbol-source": 95,
+        source: 80,
+        test: 55,
+        docs: 25,
+        lockfile: 5,
+        "low-value": 5,
+        other: 20,
+    };
+    return scores[bucket];
+}
+function genericBucketScore(bucket) {
+    const scores = {
+        "canonical-metadata": 80,
+        "overview-doc": 65,
+        "exact-path": 90,
+        "symbol-source": 85,
+        source: 60,
+        test: 45,
+        docs: 50,
+        lockfile: 5,
+        "low-value": 5,
+        other: 20,
+    };
+    return scores[bucket];
+}
+function pathTermBonus(scopePath, terms) {
+    const path = normalizedPath(scopePath);
+    const name = basename(path);
+    let score = 0;
+    for (const term of terms) {
+        if (name === term || name.startsWith(`${term}.`)) {
+            score = Math.max(score, 25);
+        }
+        else if (path.includes(term)) {
+            score = Math.max(score, 12);
+        }
+    }
+    return score;
+}
+function depthPenalty(scopePath) {
+    return Math.min(pathSegments(scopePath).length, 12);
+}
+function priorityForFile(file, terms, policy) {
+    return (bucketScore(bucketByPath(file.relativePath), policy.intent) +
+        pathTermBonus(file.relativePath, terms) -
+        depthPenalty(file.relativePath));
+}
+function bucketCounts(files) {
+    const counts = emptyBucketCounts();
+    for (const file of files) {
+        counts[bucketByPath(file.relativePath)] += 1;
+    }
+    return counts;
+}
+function sortFiles(files, query, policy) {
+    // Score each file ONCE (query tokenization + path bucketing are O(path) and were previously
+    // recomputed twice per comparison — O(n log n) blocking work on the 2000-file candidate cap).
+    // Tie-break on raw code-point order, not localeCompare, so evidence ordering is reproducible
+    // across locales/ICU builds (regulated-delivery determinism).
+    const terms = normalizedQueryTerms(query);
+    const scored = files.map((file) => ({ file, score: priorityForFile(file, terms, policy) }));
+    scored.sort((a, b) => {
+        if (a.score !== b.score)
+            return b.score - a.score;
+        if (a.file.relativePath < b.file.relativePath)
+            return -1;
+        return a.file.relativePath > b.file.relativePath ? 1 : 0;
+    });
+    return scored.map((entry) => entry.file);
+}
+export function resolveSearchPolicy(hasExplicitRelativePaths, hints) {
+    const mode = hasExplicitRelativePaths ? "explicit-scope" : "workspace-root-default";
+    const intent = hints?.retrievalIntent ?? "generic";
+    return {
+        mode,
+        intent,
+        applyGitignore: mode === "workspace-root-default",
+        omitLowValueWorkspaceFiles: mode === "workspace-root-default",
+    };
+}
+// The legacy `gatherCandidates(scope, limits, fs)` overload predates the search-policy work and is
+// still used by importGraph / testSourcePairing, which expect FULL-tree discovery (no .gitignore
+// subset, no low-value omission) — that was the hardcoded `applyGitignore: false` behavior before
+// this change. Resolving the default workspace-root policy for them would silently start dropping
+// gitignored / low-value files from those callers, so they get this neutral policy instead.
+export function legacyDiscoveryPolicy(hasExplicitRelativePaths) {
+    return {
+        mode: hasExplicitRelativePaths ? "explicit-scope" : "workspace-root-default",
+        intent: "generic",
+        applyGitignore: false,
+        omitLowValueWorkspaceFiles: false,
+    };
+}
+export function policyOmissionReason(scopePath, policy) {
+    if (!policy.omitLowValueWorkspaceFiles) {
+        return undefined;
+    }
+    if (isLockfile(scopePath) && policy.intent !== "project-metadata") {
+        return "generated";
+    }
+    return hasLowValueSegment(scopePath) ? "generated" : undefined;
+}
+export function extraIgnoreLinesForSearch(policy) {
+    return policy.omitLowValueWorkspaceFiles ? LOW_VALUE_IGNORE_LINES : [];
+}
+export function orderCandidatesForSearch(files, query, policy, ignoredByDiscovery, deniedByDiscovery) {
+    const ordered = sortFiles(files, query, policy);
+    return {
+        files: ordered,
+        diagnostics: {
+            policyMode: policy.mode,
+            intent: policy.intent,
+            filesDiscovered: files.length,
+            filesAfterPolicy: ordered.length,
+            ignoredByDiscovery,
+            deniedByDiscovery,
+            candidateBuckets: bucketCounts(ordered),
+        },
+    };
+}
+export function shouldScoreContent(query, text, policy) {
+    if (query.kind !== "natural-language" && query.kind !== "exact-symbol") {
+        return true;
+    }
+    if (policy.intent === "repository-overview" || policy.intent === "project-metadata") {
+        return true;
+    }
+    const haystack = query.caseSensitive ? text : text.toLowerCase();
+    const terms = normalizedQueryTerms(query).filter((term) => term.length >= 4);
+    return terms.length === 0 || terms.some((term) => haystack.includes(term));
+}

package/node_modules/@oscharko-dev/keiko-workspace/dist/repoSearchScan.d.ts

@@ -1,6 +1,7 @@
-import type { CandidateFile, CandidateOmissionReason, EvidenceAtom, EvidenceAtomProvenanceKind } from "@oscharko-dev/keiko-contracts/connected-context";
+import type { CandidateFile, CandidateOmissionReason, EvidenceAtom, EvidenceAtomProvenanceKind, RetrievalQuery } from "@oscharko-dev/keiko-contracts/connected-context";
 import type { WorkspaceFs } from "./fs.js";
 import type { LineMatcher } from "./repoSearchMatchers.js";
+import { type SearchDiagnostics, type SearchPolicy } from "./repoSearchPolicy.js";
 import type { DiscoveredFile, WorkspaceInfo } from "./types.js";
 export declare function isImageScopePath(scopePath: string): boolean;
 export interface ScopeShape {
@@ -32,8 +33,10 @@ export declare function buildCandidate(scopePath: string, omitted: CandidateOmis
 export interface CandidateSet {
     readonly files: readonly DiscoveredFile[];
     readonly truncated: boolean;
+    readonly diagnostics: SearchDiagnostics;
 }
 export declare function gatherCandidates(scope: ScopeShape, limits: LimitsShape, fs: WorkspaceFs): CandidateSet;
+export declare function gatherCandidates(scope: ScopeShape, query: RetrievalQuery, limits: LimitsShape, fs: WorkspaceFs, policy: SearchPolicy): CandidateSet;
 export declare function probeBinary(fs: WorkspaceFs, abs: string, size: number): Promise<boolean>;
 export interface SearchTextRunner {
     readonly scope: ScopeShape;
@@ -43,6 +46,8 @@ export interface SearchTextRunner {
     readonly startMs: number;
     readonly matcher: LineMatcher;
     readonly fingerprint: string;
+    readonly policy: SearchPolicy;
+    readonly query: RetrievalQuery;
 }
 export interface RunState {
     filesScanned: number;

package/node_modules/@oscharko-dev/keiko-workspace/dist/repoSearchScan.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"repoSearchScan.d.ts","sourceRoot":"","sources":["../src/repoSearchScan.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EACV,aAAa,EACb,uBAAuB,EACvB,YAAY,EACZ,0BAA0B,EAC3B,MAAM,iDAAiD,CAAC;AAOzD,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAQ3C,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,yBAAyB,CAAC;AAC3D,OAAO,KAAK,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAsBhE,wBAAgB,gBAAgB,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAK3D;AAED,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,SAAS,EAAE,aAAa,CAAC;IAClC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,aAAa,EAAE,SAAS,MAAM,EAAE,CAAC;CAC3C;AAED,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC;IACjC,QAAQ,CAAC,kBAAkB,EAAE,MAAM,CAAC;IACpC,QAAQ,CAAC,sBAAsB,EAAE,MAAM,CAAC;IACxC,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAC;CAC/B;AAED,MAAM,WAAW,SAAS;IACxB,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,QAAQ,CAAC,SAAS,EAAE;QAAE,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;QAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAAC;IACzF,QAAQ,CAAC,cAAc,EAAE,0BAA0B,CAAC;IACpD,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,gBAAgB,EAAE,MAAM,CAAC;IAClC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;CAC9B;AAED,wBAAgB,SAAS,CAAC,KAAK,EAAE,SAAS,GAAG,YAAY,CAwBxD;AAED,wBAAgB,cAAc,CAC5B,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,uBAAuB,GAAG,SAAS,GAC3C,aAAa,CAEf;AAkBD,MAAM,WAAW,YAAY;IAC3B,QAAQ,CAAC,KAAK,EAAE,SAAS,cAAc,EAAE,CAAC;IAC1C,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAC;CAC7B;AAED,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE,UAAU,EACjB,MAAM,EAAE,WAAW,EACnB,EAAE,EAAE,WAAW,GACd,YAAY,CAuBd;AAED,wBAAsB,WAAW,CAAC,EAAE,EAAE,WAAW,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAU9F;AAED,MAAM,WAAW,gBAAgB;IAC/B,QAAQ,CAAC,KAAK,EAAE,UAAU,CAAC;IAC3B,QAAQ,CAAC,MAAM,EAAE,WAAW,CAAC;IAC7B,QAAQ,CAAC,EAAE,EAAE,WAAW,CAAC;IACzB,QAAQ,CAAC,KAAK,EAAE,MAAM,MAAM,CAAC;IAC7B,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC;IAC9B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;CAC9B;AAED,MAAM,WAAW,QAAQ;IACvB,YAAY,EAAE,MAAM,CAAC;IACrB,eAAe,EAAE,MAAM,CAAC;IACxB,SAAS,EAAE,OAAO,CAAC;CACpB;AAED,wBAAgB,OAAO,CAAC,MAAM,EAAE,gBAAgB,GAAG,MAAM,CAExD;AAED,wBAAgB,QAAQ,CAAC,MAAM,EAAE,gBAAgB,EAAE,KAAK,EAAE,QAAQ,GAAG,OAAO,CAc3E;AAgBD,wBAAgB,SAAS,CAAC,GAAG,EAAE,OAAO,GAAG,OAAO,CAM/C;AAkED,wBAAsB,QAAQ,CAC5B,MAAM,EAAE,gBAAgB,EACxB,IAAI,EAAE,cAAc,EACpB,KAAK,EAAE,QAAQ,EACf,KAAK,EAAE,YAAY,EAAE,EACrB,UAAU,EAAE,aAAa,EAAE,GAC1B,OAAO,CAAC,IAAI,CAAC,CAqCf"}
+{"version":3,"file":"repoSearchScan.d.ts","sourceRoot":"","sources":["../src/repoSearchScan.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EACV,aAAa,EACb,uBAAuB,EACvB,YAAY,EACZ,0BAA0B,EAC1B,cAAc,EACf,MAAM,iDAAiD,CAAC;AAOzD,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAQ3C,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,yBAAyB,CAAC;AAC3D,OAAO,EAOL,KAAK,iBAAiB,EACtB,KAAK,YAAY,EAClB,MAAM,uBAAuB,CAAC;AAC/B,OAAO,KAAK,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAsBhE,wBAAgB,gBAAgB,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAK3D;AAED,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,SAAS,EAAE,aAAa,CAAC;IAClC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,aAAa,EAAE,SAAS,MAAM,EAAE,CAAC;CAC3C;AAED,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC;IACjC,QAAQ,CAAC,kBAAkB,EAAE,MAAM,CAAC;IACpC,QAAQ,CAAC,sBAAsB,EAAE,MAAM,CAAC;IACxC,QAAQ,CAAC,YAAY,EAAE,MAAM,CAAC;CAC/B;AAED,MAAM,WAAW,SAAS;IACxB,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,QAAQ,CAAC,SAAS,EAAE;QAAE,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;QAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAAC;IACzF,QAAQ,CAAC,cAAc,EAAE,0BAA0B,CAAC;IACpD,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,gBAAgB,EAAE,MAAM,CAAC;IAClC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IACvB,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;CAC9B;AAED,wBAAgB,SAAS,CAAC,KAAK,EAAE,SAAS,GAAG,YAAY,CAwBxD;AAED,wBAAgB,cAAc,CAC5B,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,uBAAuB,GAAG,SAAS,GAC3C,aAAa,CAEf;AA+BD,MAAM,WAAW,YAAY;IAC3B,QAAQ,CAAC,KAAK,EAAE,SAAS,cAAc,EAAE,CAAC;IAC1C,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAC;IAC5B,QAAQ,CAAC,WAAW,EAAE,iBAAiB,CAAC;CACzC;AA4CD,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE,UAAU,EACjB,MAAM,EAAE,WAAW,EACnB,EAAE,EAAE,WAAW,GACd,YAAY,CAAC;AAChB,wBAAgB,gBAAgB,CAC9B,KAAK,EAAE,UAAU,EACjB,KAAK,EAAE,cAAc,EACrB,MAAM,EAAE,WAAW,EACnB,EAAE,EAAE,WAAW,EACf,MAAM,EAAE,YAAY,GACnB,YAAY,CAAC;AA2ChB,wBAAsB,WAAW,CAAC,EAAE,EAAE,WAAW,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAU9F;AAED,MAAM,WAAW,gBAAgB;IAC/B,QAAQ,CAAC,KAAK,EAAE,UAAU,CAAC;IAC3B,QAAQ,CAAC,MAAM,EAAE,WAAW,CAAC;IAC7B,QAAQ,CAAC,EAAE,EAAE,WAAW,CAAC;IACzB,QAAQ,CAAC,KAAK,EAAE,MAAM,MAAM,CAAC;IAC7B,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC;IAC9B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;IAC7B,QAAQ,CAAC,MAAM,EAAE,YAAY,CAAC;IAC9B,QAAQ,CAAC,KAAK,EAAE,cAAc,CAAC;CAChC;AAED,MAAM,WAAW,QAAQ;IACvB,YAAY,EAAE,MAAM,CAAC;IACrB,eAAe,EAAE,MAAM,CAAC;IACxB,SAAS,EAAE,OAAO,CAAC;CACpB;AAED,wBAAgB,OAAO,CAAC,MAAM,EAAE,gBAAgB,GAAG,MAAM,CAExD;AAED,wBAAgB,QAAQ,CAAC,MAAM,EAAE,gBAAgB,EAAE,KAAK,EAAE,QAAQ,GAAG,OAAO,CAc3E;AAgBD,wBAAgB,SAAS,CAAC,GAAG,EAAE,OAAO,GAAG,OAAO,CAM/C;AAkED,wBAAsB,QAAQ,CAC5B,MAAM,EAAE,gBAAgB,EACxB,IAAI,EAAE,cAAc,EACpB,KAAK,EAAE,QAAQ,EACf,KAAK,EAAE,YAAY,EAAE,EACrB,UAAU,EAAE,aAAa,EAAE,GAC1B,OAAO,CAAC,IAAI,CAAC,CA0Cf"}

package/node_modules/@oscharko-dev/keiko-workspace/dist/repoSearchScan.js

@@ -3,7 +3,7 @@
 // Every file system touch goes through the injected WorkspaceFs port; nothing here calls
 // node:fs directly.
 import { CONNECTED_CONTEXT_SCHEMA_VERSION, isValidScopePath, } from "@oscharko-dev/keiko-contracts/connected-context";
-import { discoverFiles, readWorkspaceFile } from "./discovery.js";
+import { discoverWithStats, readWorkspaceFile } from "./discovery.js";
 import { FileTooLargeError, RepoSearchInvalidQueryError } from "./errors.js";
 import { isDenied } from "./ignore.js";
 import { resolveWithinWorkspace } from "./paths.js";
@@ -12,6 +12,7 @@ import { looksBinary } from "./binaryDetect.js";
 import { collectFromEntries } from "./repoSearchEntries.js";
 import { collectBestLines } from "./repoSearchLineSelection.js";
 import { evidenceAtomStableId } from "./stableId.js";
+import { extraIgnoreLinesForSearch, legacyDiscoveryPolicy, orderCandidatesForSearch, policyOmissionReason, resolveSearchPolicy, shouldScoreContent, } from "./repoSearchPolicy.js";
 const BINARY_PROBE_BYTES = 512;
 const IMAGE_EXTENSIONS = new Set([
     ".avif",
@@ -64,14 +65,52 @@ export function buildAtom(shape) {
 export function buildCandidate(scopePath, omitted) {
     return { scopePath, score: 0, signals: [], omitted };
 }
-function collectFromDirectory(scope, limits, fs) {
-    const files = discoverFiles(scope.workspace, { maxDepth: 12, maxFiles: limits.maxFilesScanned + 1, applyGitignore: false }, fs);
+function collectFromDirectory(scope, limits, fs, policy) {
+    const extraIgnoreLines = extraIgnoreLinesForSearch(policy);
+    const workspace = extraIgnoreLines.length === 0
+        ? scope.workspace
+        : { ...scope.workspace, ignoreLines: [...scope.workspace.ignoreLines, ...extraIgnoreLines] };
+    const result = discoverWithStats(workspace, {
+        maxDepth: 12,
+        maxFiles: limits.maxFilesScanned + 1,
+        applyGitignore: policy.applyGitignore,
+    }, fs);
+    const files = result.files;
     return {
         files: files.slice(0, limits.maxFilesScanned),
         truncated: files.length > limits.maxFilesScanned,
+        ignored: result.stats.ignored,
+        denied: result.stats.denied,
     };
 }
-export function gatherCandidates(scope, limits, fs) {
+const DEFAULT_GATHER_QUERY = {
+    kind: "natural-language",
+    text: "generic repository search",
+    caseSensitive: false,
+    maxResults: 100,
+    emittedAtMs: 0,
+};
+function isRetrievalQuery(value) {
+    return typeof value === "object" && value !== null && "kind" in value && "text" in value;
+}
+function resolveGatherInputs(scope, queryOrLimits, limitsOrFs, fsOrPolicy, policy) {
+    if (isRetrievalQuery(queryOrLimits)) {
+        return {
+            query: queryOrLimits,
+            limits: limitsOrFs,
+            fs: fsOrPolicy,
+            policy: policy ?? resolveSearchPolicy(scope.relativePaths.length > 0, undefined),
+        };
+    }
+    return {
+        query: DEFAULT_GATHER_QUERY,
+        limits: queryOrLimits,
+        fs: limitsOrFs,
+        policy: legacyDiscoveryPolicy(scope.relativePaths.length > 0),
+    };
+}
+export function gatherCandidates(scope, queryOrLimits, limitsOrFs, fsOrPolicy, policy) {
+    const inputs = resolveGatherInputs(scope, queryOrLimits, limitsOrFs, fsOrPolicy, policy);
     // Defense in depth alongside the realpath gate: validate scope.relativePaths against the
     // contracts-layer shape rules (no absolute paths, no `..`, no drive letters, no backslashes).
     // resolveWithinWorkspace + assertContainedRealPath already provide a complete barrier; this
@@ -83,16 +122,20 @@ export function gatherCandidates(scope, limits, fs) {
         }
     }
     if (scope.relativePaths.length === 0) {
-        const result = collectFromDirectory(scope, limits, fs);
+        const result = collectFromDirectory(scope, inputs.limits, inputs.fs, inputs.policy);
+        const ordered = orderCandidatesForSearch(result.files, inputs.query, inputs.policy, result.ignored, result.denied);
         return {
-            files: [...result.files].sort((a, b) => (a.relativePath < b.relativePath ? -1 : 1)),
+            files: ordered.files,
             truncated: result.truncated,
+            diagnostics: ordered.diagnostics,
         };
     }
-    const result = collectFromEntries(scope, limits, fs);
+    const result = collectFromEntries(scope, inputs.limits, inputs.fs);
+    const ordered = orderCandidatesForSearch(result.files, inputs.query, inputs.policy, 0, 0);
     return {
-        files: [...result.files].sort((a, b) => (a.relativePath < b.relativePath ? -1 : 1)),
+        files: ordered.files,
         truncated: result.truncated,
+        diagnostics: ordered.diagnostics,
     };
 }
 export async function probeBinary(fs, abs, size) {
@@ -199,6 +242,11 @@ export async function scanFile(runner, file, state, atoms, candidates) {
         candidates.push(buildCandidate(file.relativePath, "ignored"));
         return;
     }
+    const omitted = policyOmissionReason(file.relativePath, runner.policy);
+    if (omitted !== undefined) {
+        candidates.push(buildCandidate(file.relativePath, omitted));
+        return;
+    }
     let isBinary;
     try {
         isBinary = await probeBinary(runner.fs, contained.path, file.sizeBytes);
@@ -217,7 +265,7 @@ export async function scanFile(runner, file, state, atoms, candidates) {
     }
     state.filesScanned += 1;
     const text = readForScan(runner, file.relativePath, candidates);
-    if (text === undefined) {
+    if (text === undefined || !shouldScoreContent(runner.query, text, runner.policy)) {
         return;
     }
     scanLines(runner, file.relativePath, text, state, atoms);

package/node_modules/@oscharko-dev/keiko-workspace/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@oscharko-dev/keiko-workspace",
-  "version": "0.2.0-beta.7",
+  "version": "0.2.0-beta.9",
   "private": true,
   "type": "module",
   "license": "Apache-2.0",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@oscharko-dev/keiko",
-  "version": "0.2.0-beta.7",
+  "version": "0.2.0-beta.9",
   "description": "Enterprise model-agnostic developer-assist coding agent for regulated engineering workflows.",
   "type": "module",
   "license": "Apache-2.0",
@@ -45,6 +45,7 @@
     "test:coverage:ui": "vitest run --config packages/keiko-ui/vitest.coverage.config.ts --coverage",
     "test:coverage:quality": "npm run test:coverage:packages && npm run test:coverage:ui && npm run check:coverage && npm run check:coverage:release-targets && npm run check:coverage:branches",
     "test:e2e:smoke": "playwright test --project=chromium --grep @smoke",
+    "test:e2e:local-knowledge": "playwright test --config playwright.local-knowledge-regression.config.ts",
     "test:release:smoke": "npm run test:e2e:smoke",
     "dev:start": "node scripts/dev-start.mjs",
     "dev:stop": "node scripts/dev-stop.mjs",